The way clinical researchers measure patient health is undergoing a fundamental shift. For decades, the gold standard has been episodic measurement: periodic blood draws, intermittent vital signs, in-clinic electrocardiograms (ECGs), and the occasional patient-reported outcome. These snapshots, taken days or weeks apart, have long been assumed to paint an adequate picture of how a patient responds to an investigational therapy.
But emerging evidence suggests otherwise. A growing body of research indicates that critical physiological events—arrhythmias, drug-induced cardiac changes, recovery patterns—often occur in the gaps between those snapshots. A recent study from Brigham and Women’s Hospital, for instance, found that 24% of cardiac surgery patients who developed post-discharge atrial fibrillation never showed any sign of it during hospitalization, with many episodes lasting less than five minutes, too brief to be reliably captured by standard in-clinic ECGs.
This realization is driving a transformation in how sponsors, contract research organizations (CROs), and regulators think about evidence generation. Continuous physiologic monitoring—using medical-grade wearable sensors to capture heart rhythm, activity patterns, and other biometric signals on an ongoing basis—is moving from a niche tool for early-stage feasibility studies to a core component of late-stage drug development.
To understand this evolution, Drug and Device World spoke with Jiang Li, PhD, CEO of Vivalink, a digital healthcare solutions provider whose technology is being used in post-operative AFib studies, high-risk pregnancy arrhythmia research, and oncology trials. Li, who has been thinking about this space not as a gadget trend but as a measurement evolution, explains how continuous signals can fill the gaps left by episodic measurements—and what needs to happen before these signals can be trusted as regulatory-grade endpoints.
What Traditional Trials Miss
The fundamental limitation of traditional clinical trial monitoring is its episodic nature. Patients come to a site, measurements are taken, and they leave—sometimes not to return for weeks. In between, critical physiological changes can occur undetected.
“In some drug cases, it’s about safety signals,” Li explains during the interview. “After a patient takes certain medications, there could be different reactions to the drug. It’s not sufficient to take the measurements a few days or a couple of weeks after taking the drug. If the data can be taken continuously post-drug-taking, that will paint a much better picture about what’s going on with the patient.”
This is particularly relevant in cardiac safety monitoring. Traditional approaches rely on periodic in-clinic ECGs, which can miss arrhythmias or drug-induced cardiac changes that occur outside those limited windows. The consequence is a problematic blind spot: important signals of conditions like atrial fibrillation (AFib) may go undetected, potentially compromising both patient safety and trial integrity.
The AFib case is instructive. “A lot of times, in the case of rhythm management drugs, you want to try to control AFib burden for the patient,” Li notes. “To understand how the burden changes over time, you paint a much more complete picture if you have continuous data. Similarly, some drugs could cause AFib—in some severe cases, diabetes management drugs could cause AFib. If you don’t monitor the patient in a much more frequent fashion, you may just miss it.”
This dual role—enabling both efficacy assessment and safety detection—is what makes continuous monitoring particularly valuable. In Li’s assessment, “In the example of AFib, you can see both. Conditional monitoring can help the efficacy statement as well as the safety statement.”
Beyond Consumer Gadgets
As wearable technology proliferates in the consumer market—with Fitbit, Apple Watch, and similar devices now ubiquitous—a critical distinction emerges between consumer-grade and regulatory-grade data. Not all wearables are created equal, and the difference matters enormously when the data will support a regulatory submission.
Li is unequivocal on this point. “The simplest way is whether it’s really got the FDA or CE mark blessing, the regulatory clearance. Once they go through regulatory clearance, it fulfills the regulatory bodies’ requirements as far as data validation.”
But clearance alone isn’t sufficient. The specific claims and conditions of that clearance are equally important. “Look at the details about those filings,” Li advises. “For heart rate measurement, is it filed for use in an ambulatory condition? Or maybe some heart rate monitors only say they achieve certain accuracy when the patient is sitting still. That’s very different. You look at the conditions under which they are cleared, and make sure they match whatever the drug manufacturers intend to use or intend to capture the data with.”
This nuance matters because the context of data collection in a clinical trial is fundamentally different from consumer use. A patient in a trial may be active, may be experiencing drug effects, and may—intentionally or unintentionally—introduce artifacts into the data. A sensor that works well for a stationary user may fail to capture meaningful data from a patient going about daily life.
Li acknowledges the challenge: “The outliers need to be dealt with and properly managed. That’s why a tool that really qualifies for clinical trials needs to have the rigor about data quality and data accuracy, so it can be trusted by clinical trial sponsors.”
From Algorithm to Endpoint
For continuous monitoring data to be accepted as clinical trial endpoints, a clear regulatory pathway must be followed. The FDA has published specific guidelines for using digital biomarkers, emphasizing that these measures must be “very well validated and fit for purpose”.
“All the digital biomarkers need to be validated with sufficient patient population and specific usage cases for the drug manufacturers,” Li explains. “For arrhythmia or AFib monitoring, sufficient data needs to be submitted to the FDA for those algorithms to be deemed sufficient for AFib burden measurement. All those need to be aligned and cleared with the FDA before they can be used as digital endpoints.”
This validation process requires demonstrating not only the accuracy of the sensor but also the clinical relevance of the measures it produces. The FDA has signaled growing support for digital health technologies in later-stage trials, noting that wearable sensors and remote monitoring tools can make it easier to collect frequent or continuous data from participants, including data that might not show up during a typical site visit. The agency has also recognized that these technologies can help researchers detect novel endpoints and gather more granular insights than conventional methods.
The stakes are high, but so are the potential returns. As Li notes, continuous data can enable more efficient trials, potentially reducing sample sizes while maintaining statistical power. “In engineering terms, with a much larger volume of continuous data, you have more statistical power. The number of participants can be reduced—in our assessment, with continuous data, you could get to anywhere from 20 to 30% reduction in population.”
Three Operational Challenges—and How to Address Them
Despite the clear benefits, integrating continuous monitoring into clinical trials is not without hurdles. Li identifies three major challenges that sponsors and CROs must navigate.
The first is regulatory risk. “Making sure adding biosensor data does not cause any complication or risk for the sponsor’s FDA submission,” Li says. “If they’re not using a good choice of sensor—maybe it’s not intended for the intended use case for the trial—then there’s increased risk about using those data. They have to be very careful about selecting what kind of tool they use.”
The second challenge is operational. Global trials, which may involve hundreds of sites across multiple countries, introduce logistical complexity. “It’s complicated logistics, multi-country data management and data governance. That’s another factor.”
The third is the budget. “Very often, sponsors are concerned about their budget,” Li acknowledges. “They wonder: ‘I’m using a new tool, a new method—will this blow out my budget?’”
But Li argues that with the right approach, the return on investment justifies the initial outlay. “If we talk about potentially reducing the number of participants in the trial, and higher quality data that help the pharma company prepare for the one-trial default, all these are really good returns. With the right partner, you can really control the risk.”
The One-Trial Default and the Push for Higher Quality Data
Li references a recent FDA development that underscores the growing importance of high-quality, continuous data: the agency’s move toward a “one-trial default.” Rather than requiring two pivotal trials for drug approval, the FDA is pivoting toward accepting a single, well-designed trial—but with higher expectations for data quality.
“From our understanding, this will push the sponsor to have higher quality data in their submission package to justify the one-trial position,” Li explains. “Continuous sensor data would help them justify that position better.”
This aligns with broader regulatory trends. Recent FDA guidance (2024) and EMA frameworks emphasize participant safety and data integrity as core principles, requiring clear remote monitoring protocols and validation of digital systems. The ICH E6(R3) guidelines, currently in development, are expected to further reinforce risk-based monitoring approaches that leverage continuous data streams.
Real-World Evidence: AFib, Pregnancy, and Beyond
The value of continuous monitoring is not merely theoretical. Real-world studies are demonstrating their practical utility across therapeutic areas.
In the cardiovascular space, researchers at the University of California, San Francisco (UCSF) are running a multi-year hybrid study on AFib using Vivalink’s wearable ECG patch to collect continuous heart rhythm data. Early signals in the wearable data helped the team identify new paths worth exploring, including a related substudy focused on CHIP (clonal hematopoiesis of indeterminate potential), a blood cell mutation linked to AFib and inflammation, and a separate study tracking heart rhythm changes in high-risk pregnant patients.
The HEARTBEAT study, a large-scale decentralized trial exploring the use of smartwatches to collect biometric data for cardiovascular disease research, demonstrates the scalability of these approaches. With 10,000 adult participants being remotely monitored over a one-year period, the study has shown that a fully decentralized approach can be executed at scale in a highly regulated environment.
In neurology, Vivalink is supporting a hospital-based study at Ochsner Health exploring the link between seizures and cardiac function. Early findings suggest seizures may disrupt heart function by interfering with coordination between the heart’s upper and lower chambers, potentially reducing oxygen flow to the brain. Some patients exhibit irregular heart rhythms during seizures, which could heighten the risk of sudden unexpected death in epilepsy (SUDEP). Identifying these patterns may help clinicians identify high-risk patients and develop targeted interventions.
Home-Based Healthcare and Continuous Evidence
Looking ahead, Li sees the convergence of clinical trials and patient care as an inevitable trend. “We have seen more and more cases that patient care is moving away from the hospital setting, happening at home. It’s inevitable that drug administration is going to happen more at home as well. With that setting, home sensor data for understanding patient reaction to drugs is a natural trend.”
One example Li cites is in oncology. “Some of the latest oncology drugs have high efficacy but also high concerns about drug toxicity. They are limited to administration only in hospital settings, not able to move into clinic or home settings. By deploying continuous monitoring tools and more capable digital biomarkers, it’s very possible that these drugs can move into a home setting with more reliable remote monitoring technology enabled.”
This vision aligns with broader industry trends. Decentralized and hybrid clinical trial models are reshaping clinical research, leveraging digital health technologies to improve patient access, enhance data quality, and optimize operational efficiency. Hybrid trials—combining remote elements like telehealth and wearables with targeted in-person visits—have become the preferred approach for many sponsors, balancing innovation with regulatory compliance.
A Nascent Technology with Growing Momentum
Li acknowledges that continuous monitoring is still nascent, particularly for the clinical trial industry. But the momentum is building.
“The good news is that a lot of people are starting to see the value of it. A very good example of one such success case is CGM [continuous glucose monitoring]. Now it’s becoming mainstream—you can get CGM monitoring from Walgreens. That’s one good example, but there will be more and more of this kind of capability coming into play. It will enable both better clinical trial capabilities and patient care capabilities.”
For sponsors navigating this transition, the path forward involves careful selection of validated, regulatory-cleared sensors; thoughtful integration with existing data systems; and a clear understanding of how continuous data can support both regulatory approval and post-market value demonstration to payers.
“The price for drugs really needs to show more evidence about how effective the drug is in real life and how the quality of life of a patient is going to be affected,” Li notes. “All these are going to be very important assessments for patients and for drug usage.”
As the industry moves toward this future, the question is no longer whether continuous monitoring will play a role in clinical trials, but how quickly sponsors will adapt to a world where the snapshot is no longer enough.
